CN109276255B - Method and device for detecting tremor of limbs - Google Patents
Method and device for detecting tremor of limbs Download PDFInfo
- Publication number
- CN109276255B CN109276255B CN201811422696.1A CN201811422696A CN109276255B CN 109276255 B CN109276255 B CN 109276255B CN 201811422696 A CN201811422696 A CN 201811422696A CN 109276255 B CN109276255 B CN 109276255B
- Authority
- CN
- China
- Prior art keywords
- limb
- sampling frequency
- recognition model
- preset
- sensor
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
- A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb
- A61B5/1101—Detecting tremor
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
- A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Physiology (AREA)
- Veterinary Medicine (AREA)
- Public Health (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Dentistry (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Psychiatry (AREA)
- Signal Processing (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The embodiment of the invention provides a method and a device for detecting tremor of limbs, which relate to the technical field of artificial intelligence and comprise the following steps: constructing and training an identification model, wherein the identification model is used for identifying whether the limb signals have tremor characteristics; acquiring a to-be-identified limb signal of a first user, which is acquired by a sensor in a mobile terminal; inputting the limb signal to be recognized into the trained recognition model; and the obtained recognition model recognizes the limb signal to be recognized according to the trained network parameters, outputs a recognition result and sends the recognition result to the mobile terminal. The technical scheme provided by the embodiment of the invention can solve the problem of low accuracy of the limb tremor detection in the prior art.
Description
[ technical field ] A method for producing a semiconductor device
The invention relates to the technical field of artificial intelligence, in particular to a method and a device for detecting tremor of limbs.
[ background of the invention ]
At present, as the population ages gradually, some people suffering from hyperthyroidism, parkinson or other special diseases can have the symptoms of tremor of the limbs such as hands, heads and legs, and inconvenience is brought to daily life, work, social interaction and the like of patients. The existing limb tremor detection method generally needs to detect through professional medical detection equipment or through some auxiliary detection equipment, and the detection result is low in accuracy.
Therefore, how to improve the accuracy of detecting tremor of limbs is an urgent problem to be solved at present.
[ summary of the invention ]
In view of this, embodiments of the present invention provide a method and an apparatus for detecting tremor of limbs, so as to solve the problem of low accuracy in detecting tremor of limbs in the prior art.
In order to achieve the above object, according to one aspect of the present invention, there is provided a limb tremor detection method, the method including:
constructing and training an identification model, wherein the identification model is used for identifying whether the limb signals have tremor characteristics; acquiring a to-be-identified limb signal of a first user, which is acquired by a sensor in a mobile terminal; inputting the limb signal to be recognized into the trained recognition model; and acquiring the recognition model, recognizing the limb signal to be recognized according to the trained network parameters, outputting a recognition result, and sending the recognition result to the mobile terminal.
Further, the building and training of the recognition model comprises: acquiring first limb signals of a plurality of healthy human bodies and second limb signals of a plurality of limb tremor symptom patients; respectively carrying out sample preparation on the plurality of first limb signals and the plurality of second limb signals according to a preset format to obtain a training set comprising a plurality of training samples; constructing the recognition model; inputting the training set into a convolutional neural network of the recognition model to obtain the forward output of the convolutional neural network; and updating the weight and the bias of the convolutional neural network by using a back propagation neural network algorithm to obtain the trained recognition model, and storing the network parameters of the trained recognition model.
Further, the method for obtaining the trained recognition model by updating the weight and the bias of the convolutional neural network by using a back propagation neural network algorithm includes:
constructing a loss function according to the forward output and the real result of the training sample, wherein the expression of the loss function is
Wherein, E loss Representing the loss function, n representing the total number of training samples, x i Representing the forward output of the i-th training sample, y i Is represented by x i The true result of the corresponding ith training sample; and updating the weight and the bias of the convolutional neural network by adopting a back propagation algorithm based on batch gradient descent, and optimizing the loss function to minimize the loss function to obtain the trained recognition model.
Further, after acquiring the to-be-recognized limb signal of the first user acquired by the sensor in the mobile terminal and before inputting the to-be-recognized limb signal into the trained recognition model, the method further includes:
and preprocessing the limb signal to be identified, wherein the preprocessing comprises filtering a low-frequency human motion signal by adopting a Kalman filtering algorithm.
Further, the sensor comprises at least one of a multi-axial acceleration sensor, a multi-axial gyroscope, and a multi-axial inclinometer; the mobile terminal is any one of a mobile phone, an iPad, a smart watch or a wearable smart device.
Further, before acquiring the to-be-identified limb signal of the first user acquired by the sensor in the mobile terminal, the method further includes:
setting a preset sampling frequency of the sensor to enable the sensor to sample according to the preset sampling frequency to obtain actual sampling data with preset sampling duration; calculating the actual sampling frequency of the sensor according to the actual sampling data; judging whether a frequency error value between the actual sampling frequency and the preset sampling frequency exceeds a preset error range or not; when the frequency error value is within the preset error range, gradually increasing the preset sampling frequency until the measured frequency error value between the actual sampling frequency and the preset sampling frequency exceeds the preset error range, and taking the current preset sampling frequency as the maximum sampling frequency of the sensor; and acquiring the to-be-identified limb signal of the first user at the maximum sampling frequency.
In order to achieve the above object, according to one aspect of the present invention, there is provided a limb tremor detection apparatus, the apparatus including: the system comprises a construction unit, a training unit and a recognition unit, wherein the construction unit is used for constructing and training a recognition model, and the recognition model is used for recognizing whether a limb signal has tremor characteristics; the system comprises an acquisition unit, a processing unit and a processing unit, wherein the acquisition unit is used for acquiring a to-be-identified limb signal of a first user, which is acquired by a sensor in the mobile terminal; the input unit is used for inputting the limb signal to be recognized into the trained recognition model; and the sending unit is used for acquiring the recognition model, recognizing the limb signal to be recognized according to the trained network parameters, outputting a recognition result and sending the recognition result to the mobile terminal.
Further, the construction unit includes: the acquiring subunit is used for acquiring first limb signals of a plurality of healthy human bodies and second limb signals of a plurality of limb tremor symptom patients; the manufacturing subunit is used for respectively manufacturing samples of the plurality of first limb signals and the plurality of second limb signals according to a preset format to obtain a training set comprising a plurality of training samples; a construction subunit configured to construct the recognition model; the training subunit is used for inputting the training set into the convolutional neural network of the recognition model to obtain the forward output of the convolutional neural network; and the updating subunit is used for updating the weight and the bias of the convolutional neural network by using a back propagation neural network algorithm to obtain the trained recognition model and storing the network parameters of the trained recognition model.
In order to achieve the above object, according to one aspect of the present invention, there is provided a computer non-volatile storage medium including a stored program which, when executed, controls an apparatus in which the storage medium is located to perform the above-described limb tremor detection method.
To achieve the above object, according to one aspect of the present invention, there is provided a server comprising a memory for storing information including program instructions and a processor for controlling the execution of the program instructions, which are loaded and executed by the processor, to implement the steps of the limb tremor detection method described above.
In the scheme, the limb signals to be identified, which are acquired by a sensor in the mobile terminal of the user, are acquired, and the identification model for deep learning identifies the limb signals to be identified so as to judge whether the identified limb signals have tremor characteristics or not, so as to remind the user whether the tremor symptoms of the limbs exist or not. The recognition model through deep learning recognizes the limb signals to be recognized, the whole process is simple and quick, and the detection accuracy of the tremor of the limbs is improved.
[ description of the drawings ]
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flow chart of a method of limb tremor detection according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a limb tremor detection apparatus according to an embodiment of the present invention;
fig. 3 is a schematic diagram of a server according to an embodiment of the present invention.
[ detailed description ] A
For better understanding of the technical solutions of the present invention, the following detailed descriptions of the embodiments of the present invention are provided with reference to the accompanying drawings.
It should be understood that the described embodiments are only some embodiments of the invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The terminology used in the embodiments of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in the description of the invention and the appended claims, the singular forms "a", "an", and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
It should be understood that the term "and/or" as used herein is merely one type of association that describes an associated object, meaning that three relationships may exist, e.g., a and/or B may mean: a exists alone, A and B exist simultaneously, and B exists alone. In addition, the character "/" herein generally indicates that the former and latter related objects are in an "or" relationship.
It should be understood that although the terms first, second, third, etc. may be used to describe terminals in embodiments of the present invention, these terminals should not be limited by these terms. These terms are only used to distinguish one terminal from another. For example, a first terminal may also be referred to as a second terminal, and similarly, a second terminal may also be referred to as a first terminal, without departing from the scope of embodiments of the present invention.
The word "if" as used herein may be interpreted as "at 8230; \8230;" or "when 8230; \8230;" or "in response to a determination" or "in response to a detection", depending on the context. Similarly, the phrases "if determined" or "if detected (a stated condition or event)" may be interpreted as "when determined" or "in response to a determination" or "when detected (a stated condition or event)" or "in response to a detection (a stated condition or event)", depending on the context.
Fig. 1 is a flow chart of a method for detecting tremor of a limb according to an embodiment of the present invention, as shown in fig. 1, the method including:
and S101, constructing and training an identification model, wherein the identification model is used for identifying whether the limb signals have tremor characteristics.
Step S102, acquiring a to-be-identified limb signal of a first user, which is acquired by a sensor in the mobile terminal.
The mobile terminal comprises any one of a mobile phone, an ipad, a smart watch or a wearable smart device. The sensor comprises at least one of a multi-axial acceleration sensor, a multi-axial gyroscope, and a multi-axial inclinometer. Therefore, whether the limb tremor symptom is suffered or not can be detected only through the limb tremor signal acquired by the mobile terminal carried by the human body, and the detection can be carried out anytime and anywhere, so that the user can be helped to know the body state of the user in real time.
And step S103, inputting a limb signal to be recognized into the trained recognition model.
And step S104, acquiring the recognition model, recognizing the to-be-recognized limb signal according to the trained network parameters, outputting a recognition result, and sending the recognition result to the mobile terminal. So that the user can quickly and accurately know whether the user suffers from tremor.
In the scheme, the limb signals to be identified collected by the sensor in the mobile terminal of the user are obtained, and the identification model for deep learning identifies the limb signals to be identified so as to judge whether the identified limb signals have tremor characteristics and remind the user whether the user has limb tremor. The recognition model through deep learning recognizes the limb signals to be recognized, the whole process is simple and quick, and the detection accuracy of the tremor of the limbs is improved.
Optionally, constructing and training the recognition model comprises:
acquiring a plurality of first limb signals of a healthy human body and a plurality of second limb signals of a patient with limb tremor symptoms. In this embodiment, the first limb signal is a normal signal of a human body diagnosed as healthy, and the second limb signal is a tremor signal of a human body diagnosed with tremor symptoms of the limb. And taking two different limb signals as basic data for training the recognition model, so that the recognition model can be deeply learned.
The limb signals comprise parameters such as vibration frequency, amplitude, frequency spectrum and the like, the tremor can be divided into slow tremor (3-5 Hz) and fast tremor (6-12 Hz) according to different frequencies, the amplitude represents the amplitude of the tremor, and the frequency spectrum is a frequency distribution curve. The limb signal number can be used to characterize whether the body has limb tremor characteristics.
And respectively carrying out sample preparation on the plurality of first limb signals and the plurality of second limb signals according to a preset format to obtain a training set comprising a plurality of training samples. Specifically, the classification labeling is carried out on the limb signals (normal and pathological) through different labels.
Constructing an identification model, wherein the identification model comprises a convolutional neural network;
inputting the training set into a convolutional neural network of the recognition model to obtain the forward output of the convolutional neural network;
and updating the weight and the bias of the convolutional neural network by using a back propagation neural network algorithm to obtain a trained recognition model, and storing the network parameters of the trained recognition model.
Specifically, a loss function is constructed according to the forward output and the real result (label) of the training sample, and the expression of the loss function is
Wherein, E loss Representing a loss function, n representing the total number of training samples, x i Represents the forward output of the i-th training sample, y i Is represented by the formula i The real result of the corresponding ith training sample;
and updating the weight and the bias of the convolutional neural network by adopting a back propagation algorithm based on batch gradient descent, and optimizing the loss function to minimize the loss function to obtain a trained recognition model.
Optionally, constructing a recognition model, wherein the recognition model comprises a convolutional neural network and a long-time and short-time memory neural network;
inputting the training set into a convolutional neural network and a long-time memory neural network of the recognition model to obtain an output result of the recognition model;
and constructing a loss function by using the output result of the recognition model and the label of the limb signal, and iterating the loss function to the minimum to obtain the trained recognition model and the network parameters thereof.
Specifically, the loss function adopts a binary cross entropy loss (binary cross entropy) function, and specifically, the probability distributions of two classes in the training set are p and q, wherein p is a true distribution and q is a non-true distribution. A cross entropy loss function is used to measure the similarity between the two probability distributions. For a random variable X, the expectation of the amount of information for all possible values of it (E [ I (X) ] = log (1/p)) is called entropy.
The cross entropy of the two categories is:
wherein p (x) is the true distribution and q (x) is determined byEstimated probability of data calculation. It should be noted that the larger the value of the cross entropy, the larger the difference.
Using an Adam gradient descent method to iterate a binary cross entropy loss function to the minimum to obtain a trained recognition model; and acquiring network parameters of a convolutional neural network and a long-term memory neural network in the trained recognition model.
Optionally, after acquiring the limb signal to be recognized and before inputting the limb signal to be recognized into the trained recognition model, the method further includes:
and preprocessing the limb signal to be identified, wherein the preprocessing comprises filtering the low-frequency human motion signal by adopting a Kalman filtering algorithm.
Optionally, before inputting the tremor signal of the limb to be recognized, which is acquired by the sensor built in the mobile device, into the trained recognition model, the method further includes:
setting a preset sampling frequency of a sensor so that the sensor performs sampling according to the preset sampling frequency to obtain actual sampling data with preset sampling duration; calculating the actual sampling frequency of the sensor according to the actual sampling data; judging whether a frequency error value between the actual sampling frequency and the preset sampling frequency exceeds a preset error range or not;
when the frequency error value is within the preset error range, gradually increasing the preset sampling frequency until the measured frequency error value between the actual sampling frequency and the preset sampling frequency exceeds the preset error range, and taking the current preset sampling frequency as the maximum sampling frequency of the sensor; and acquiring the limb signal to be identified of the first user at the maximum sampling frequency. Understandably, the tremor signals of the limbs to be identified are acquired through the maximum sampling frequency of the sensor, so that the identification efficiency can be effectively improved.
An embodiment of the present invention provides a limb tremor detection apparatus, which is configured to perform the above-mentioned limb tremor detection method, and as shown in fig. 2, the apparatus includes: the device comprises a construction unit 10, an acquisition unit 20, an input unit 30 and a sending unit 40.
The construction unit 10 is used for constructing and training a recognition model, and the recognition model is used for recognizing whether the limb signals have tremor characteristics.
The acquiring unit 20 is configured to acquire a to-be-identified limb signal of a first user, which is acquired by a sensor in the mobile terminal;
the mobile terminal comprises any one of a mobile phone, an ipad, a smart watch or a wearable smart device. The sensor includes at least one of a multi-axial acceleration sensor, a multi-axial gyroscope, and a multi-axial inclinometer. Therefore, whether the user has the limb tremor symptom or not can be detected only through the limb tremor signal acquired by the mobile terminal carried by the human body, and the detection can be carried out anytime and anywhere, so that the user can be helped to know the self body state in real time.
And the input unit 30 is used for inputting the limb signal to be recognized into the trained recognition model.
And the sending unit 40 is used for acquiring the recognition model, recognizing the limb signal to be recognized according to the trained network parameters, outputting a recognition result and sending the recognition result to the mobile terminal. So that the user can quickly and accurately know whether the user suffers from tremor.
In the scheme, the limb signals to be identified, which are acquired by a sensor in the mobile terminal of the user, are acquired, and the identification model for deep learning identifies the limb signals to be identified so as to judge whether the identified limb signals have tremor characteristics or not, so as to remind the user whether the tremor symptoms of the limbs exist or not. The recognition model through deep learning recognizes the limb signals to be recognized, the whole process is simple and quick, and the accuracy of detecting the tremor of the limbs is improved.
Optionally, the building unit 10 includes an obtaining subunit, a making subunit, a first building subunit, a first training subunit, and a first updating subunit.
The acquiring subunit is used for acquiring a plurality of first limb signals of healthy human bodies and a plurality of second limb signals of patients with limb tremor symptoms.
In this embodiment, the first limb signal is a normal signal of a person diagnosed as healthy, and the second limb signal is a tremor signal of a person diagnosed with a limb tremor symptom. And taking two different limb signals as basic data for training the recognition model, so that the recognition model can be deeply learned.
The limb signals comprise parameters such as vibration frequency, amplitude, frequency spectrum and the like, the tremor can be divided into slow tremor (3-5 Hz) and fast tremor (6-12 Hz) according to different frequencies, the amplitude represents the amplitude of the tremor, and the frequency spectrum is a frequency distribution curve. The limb signal number can be used to characterize whether the body has limb tremor characteristics.
And the making subunit is used for respectively making samples of the plurality of first limb signals and the plurality of second limb signals according to a preset format to obtain a training set comprising a plurality of training samples. Specifically, the classification labeling is carried out on the limb signals (normal and pathological) through different labels.
The first construction subunit is used for constructing an identification model, and the identification model comprises a convolutional neural network;
the first training subunit is used for inputting the training set into the convolutional neural network of the recognition model to obtain the forward output of the convolutional neural network;
and the first updating subunit is used for updating the weight and the bias of the convolutional neural network by utilizing a back propagation neural network algorithm to obtain a trained recognition model and storing the network parameters of the trained recognition model.
Specifically, inputting a training set into a convolutional neural network of the recognition model to obtain the forward output of the convolutional neural network;
and updating the weight and the bias of the convolutional neural network by using a back propagation neural network algorithm to obtain a trained recognition model, and storing the network parameters of the trained recognition model.
Specifically, a loss function is constructed according to the forward output and the real result (label) of the training sample, and the expression of the loss function is
Wherein E is loss Representing a loss function, n representing the total number of training samples, x i Represents the forward output of the i-th training sample, y i Is represented by x i The real result of the corresponding ith training sample;
and updating the weight and the bias of the convolutional neural network by adopting a back propagation algorithm based on batch gradient descent, and optimizing the loss function to minimize the loss function to obtain a trained recognition model.
Optionally, the building unit 10 further includes a second building subunit, a second training subunit, and a second updating subunit;
the second construction subunit is used for constructing a recognition model, and the recognition model comprises a convolutional neural network and a long-time and short-time memory neural network;
the second training subunit is used for inputting the training set into the convolutional neural network and the long-time and short-time memory neural network of the recognition model to obtain an output result of the recognition model;
and the second updating subunit is used for constructing a loss function by using the output result of the recognition model and the label of the limb signal, and iterating the loss function to the minimum to obtain the trained recognition model and the network parameters thereof.
Specifically, the loss function adopts a binary cross entropy loss (binary cross entropy) function, specifically, the probability distributions of two classes in the training set are p and q, wherein p is a true distribution and q is a non-true distribution. A cross entropy loss function is used to measure the similarity between the two probability distributions. For a random variable X, the expectation of the amount of information for all possible values of it (E [ I (X) ] = log (1/p)) is called entropy.
The two-class cross entropy is:
where p (x) is the true distribution and q (x) is the estimated probability calculated from the data. It should be noted that the larger the value of the cross entropy, the larger the difference.
Iterating a binary cross entropy loss function to be minimized by using an Adam gradient descent method to obtain a trained recognition model; and acquiring network parameters of a convolutional neural network and a long-term memory neural network in the trained recognition model.
Optionally, the apparatus further comprises a processing unit for preprocessing the limb signal to be identified, the preprocessing including filtering the low-frequency human motion signal using a Kalman filtering algorithm.
Optionally, the device further comprises a setting unit, a calculating unit, a testing unit and a collecting unit.
The device comprises a setting unit, a processing unit and a processing unit, wherein the setting unit is used for setting a preset sampling frequency of a sensor so that the sensor performs sampling according to the preset sampling frequency to obtain actual sampling data with preset sampling duration;
the calculating unit is used for calculating the actual sampling frequency of the sensor according to the actual sampling data; judging whether a frequency error value between the actual sampling frequency and the preset sampling frequency exceeds a preset error range or not;
the testing unit is used for increasing the preset sampling frequency step by step when the frequency error value is within the preset error range until the measured frequency error value between the actual sampling frequency and the preset sampling frequency exceeds the preset error range, and taking the current preset sampling frequency as the maximum sampling frequency of the sensor;
and the acquisition unit is used for acquiring the tremor signals of the limbs to be identified of the first user at the maximum sampling frequency. Understandably, the tremor signals of the limbs to be identified are acquired through the maximum sampling frequency of the sensor, so that the identification efficiency can be effectively improved.
The embodiment of the invention provides a non-volatile storage medium of a computer, wherein the storage medium comprises a stored program, and when the program runs, equipment where the storage medium is located is controlled to execute the following steps:
constructing and training an identification model, wherein the identification model is used for identifying whether the limb signals have tremor characteristics; acquiring a to-be-identified limb signal of a first user, which is acquired by a sensor in a mobile terminal; inputting a limb signal to be recognized into the trained recognition model; and the obtained recognition model recognizes the limb signal to be recognized according to the trained network parameters, outputs a recognition result and sends the recognition result to the mobile terminal.
Optionally, the apparatus for controlling the storage medium when the program runs further performs the following steps: acquiring first limb signals of a plurality of healthy human bodies and second limb signals of a plurality of limb tremor symptom patients; respectively carrying out sample preparation on the plurality of first limb signals and the plurality of second limb signals according to a preset format to obtain a training set comprising a plurality of training samples; constructing an identification model; inputting the training set into a convolutional neural network of the recognition model to obtain the forward output of the convolutional neural network; and updating the weight and the bias of the convolutional neural network by using a back propagation neural network algorithm to obtain a trained recognition model, and storing the network parameters of the trained recognition model.
Optionally, the apparatus for controlling the storage medium when the program runs further performs the following steps: constructing a loss function according to the forward output and the real result of the training sample, wherein the expression of the loss function is
Wherein E is loss Representing a loss function, n representing the total number of training samples, x i Represents the forward output of the i-th training sample, y i Is represented by the formula i The real result of the corresponding ith training sample; and updating the weight and the bias of the convolutional neural network by adopting a back propagation algorithm based on batch gradient descent, and optimizing the loss function to minimize the loss function to obtain a trained recognition model.
Optionally, the apparatus for controlling the storage medium when the program runs further performs the following steps: and preprocessing the limb signal to be identified, wherein the preprocessing comprises filtering the low-frequency human motion signal by adopting a Kalman filtering algorithm.
Optionally, the apparatus for controlling the storage medium when the program runs further performs the following steps: setting a preset sampling frequency of the sensor so that the sensor performs sampling according to the preset sampling frequency to obtain actual sampling data with preset sampling duration; calculating the actual sampling frequency of the sensor according to the actual sampling data; judging whether a frequency error value between the actual sampling frequency and the preset sampling frequency exceeds a preset error range or not; when the frequency error value is within the preset error range, gradually increasing the preset sampling frequency until the frequency error value between the measured actual sampling frequency and the preset sampling frequency exceeds the preset error range, and taking the current preset sampling frequency as the maximum sampling frequency of the sensor; and acquiring the limb signal to be identified of the first user at the maximum sampling frequency.
An embodiment of the present invention provides a server 100, including a memory 101 and a processor 102, where the memory 101 is configured to store information including program instructions 103, and the processor 102 is configured to control execution of the program instructions 103, and the program instructions are loaded and executed by the processor to implement the following steps:
constructing and training an identification model, wherein the identification model is used for identifying whether the limb signals have tremor characteristics; acquiring a to-be-identified limb signal of a first user, which is acquired by a sensor in a mobile terminal; inputting a limb signal to be recognized into the trained recognition model; and the obtained recognition model recognizes the limb signal to be recognized according to the trained network parameters, outputs a recognition result and sends the recognition result to the mobile terminal.
Optionally, the program instructions when loaded and executed by a processor further implement the steps of: acquiring first limb signals of a plurality of healthy human bodies and second limb signals of a plurality of limb tremor symptom patients; respectively carrying out sample preparation on the plurality of first limb signals and the plurality of second limb signals according to a preset format to obtain a training set comprising a plurality of training samples; constructing an identification model; inputting the training set into a convolutional neural network of the recognition model to obtain the forward output of the convolutional neural network; updating the weight and the bias of the convolutional neural network by using a back propagation neural network algorithm to obtain a trained recognition model, and storing the network parameters of the trained recognition model.
Optionally, the program instructions when loaded and executed by a processor further implement the steps of: constructing a loss function according to the forward output and the real result of the training sample, wherein the expression of the loss function is
Wherein E is loss Representing a loss function, n representing the total number of training samples, x i Representing the forward output of the i-th training sample, y i Is represented by x i The real result of the corresponding ith training sample; updating weights of convolutional neural network by adopting back propagation algorithm based on batch gradient descentAnd biasing, optimizing the loss function to minimize the loss function, and obtaining a trained recognition model.
Optionally, the program instructions when loaded and executed by the processor further implement the steps of: and preprocessing the limb signal to be identified, wherein the preprocessing comprises filtering the low-frequency human motion signal by adopting a Kalman filtering algorithm.
Optionally, the program instructions when loaded and executed by a processor further implement the steps of: setting a preset sampling frequency of a sensor so that the sensor performs sampling according to the preset sampling frequency to obtain actual sampling data with preset sampling duration; calculating the actual sampling frequency of the sensor according to the actual sampling data; judging whether a frequency error value between the actual sampling frequency and the preset sampling frequency exceeds a preset error range or not; when the frequency error value is within the preset error range, gradually increasing the preset sampling frequency until the frequency error value between the measured actual sampling frequency and the preset sampling frequency exceeds the preset error range, and taking the current preset sampling frequency as the maximum sampling frequency of the sensor; and acquiring the limb signal to be identified of the first user at the maximum sampling frequency.
It should be noted that the terminal according to the embodiment of the present invention may include, but is not limited to, a Personal Computer (PC), a Personal Digital Assistant (PDA), a wireless handheld device, a Tablet Computer (Tablet Computer), a mobile phone, an MP3 player, an MP4 player, and the like.
It should be understood that the application may be an application program (native app) installed on the terminal, or may also be a web page program (webApp) of a browser on the terminal, which is not limited in this embodiment of the present invention.
It can be clearly understood by those skilled in the art that, for convenience and simplicity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
In the embodiments provided in the present invention, it should be understood that the disclosed system, apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the units is only one logical division, and there may be other divisions in actual implementation, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit may be implemented in the form of hardware, or in the form of hardware plus a software functional unit.
The integrated unit implemented in the form of a software functional unit may be stored in a computer readable storage medium. The software functional unit is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device) or a Processor (Processor) to execute some steps of the methods according to the embodiments of the present invention. And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and should not be taken as limiting the scope of the present invention, and any modifications, equivalents, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (8)
1. A server comprising a memory for storing information including program instructions and a processor for controlling the execution of the program instructions, wherein the program instructions when loaded and executed by the processor implement the steps comprising:
constructing and training an identification model, wherein the identification model is used for identifying whether the limb signals have tremor characteristics;
setting a preset sampling frequency of a sensor in the mobile terminal so that the sensor performs sampling according to the preset sampling frequency to obtain actual sampling data with preset sampling duration; calculating the actual sampling frequency of the sensor according to the actual sampling data;
when the frequency error value between the actual sampling frequency and the preset sampling frequency is within a preset error range, gradually increasing the preset sampling frequency until the measured frequency error value between the actual sampling frequency and the preset sampling frequency exceeds the preset error range, and taking the current preset sampling frequency as the maximum sampling frequency of the sensor;
acquiring a to-be-identified limb signal of a first user, which is acquired by the sensor in the mobile terminal at the maximum sampling frequency;
inputting the limb signal to be recognized into the trained recognition model;
and acquiring the recognition model, recognizing the limb signal to be recognized according to the trained network parameters, outputting a recognition result, and sending the recognition result to the mobile terminal.
2. The server of claim 1, wherein the program instructions, when loaded and executed by the processor, implement the steps of building and training a recognition model, comprising:
acquiring first limb signals of a plurality of healthy human bodies and second limb signals of a plurality of limb tremor symptom patients;
respectively carrying out sample preparation on the plurality of first limb signals and the plurality of second limb signals according to a preset format to obtain a training set comprising a plurality of training samples;
constructing the recognition model;
inputting the training set into a convolutional neural network of the recognition model to obtain forward output of the convolutional neural network;
updating the weight and the bias of the convolutional neural network by using a back propagation neural network algorithm to obtain the trained recognition model, and storing the network parameters of the trained recognition model.
3. The server according to claim 2, wherein the program instructions, when loaded and executed by the processor, implement the step of updating weights and biases of the convolutional neural network by using a back propagation neural network algorithm to obtain the trained recognition model, comprising:
constructing a loss function according to the forward output and the real result of the training sample, wherein the expression of the loss function is
Wherein, E loss Representing the loss function, n representing the total number of training samples, x i Representing the forward output of the i-th training sample, y i Is represented by the formula i The real result of the corresponding ith training sample;
and updating the weight and the bias of the convolutional neural network by adopting a back propagation algorithm based on batch gradient descent, and optimizing the loss function to minimize the loss function to obtain the trained recognition model.
4. The server according to any one of claims 1 to 3, wherein the program instructions, when loaded and executed by the processor, implement, after acquiring the limb signal to be recognized of the first user collected by the sensor in the mobile terminal and before inputting the limb signal to be recognized into the trained recognition model, further implement:
and preprocessing the limb signal to be identified, wherein the preprocessing comprises filtering a low-frequency human motion signal by adopting a Kalman filtering algorithm.
5. A server according to any of claims 1-3, wherein the sensor comprises at least one of a multi-axial acceleration sensor, a multi-axial gyroscope, a multi-axial inclinometer; the mobile terminal is any one of a mobile phone, an iPad, a smart watch or a wearable smart device.
6. A limb tremor detection device, the device comprising:
the system comprises a construction unit, a training unit and a recognition unit, wherein the construction unit is used for constructing and training a recognition model, and the recognition model is used for recognizing whether a limb signal has tremor characteristics;
the mobile terminal comprises an acquisition unit, a processing unit and a processing unit, wherein the acquisition unit is used for setting a preset sampling frequency of a sensor in the mobile terminal so that the sensor performs sampling according to the preset sampling frequency to obtain actual sampling data with preset sampling duration; calculating the actual sampling frequency of the sensor according to the actual sampling data; when the frequency error value between the actual sampling frequency and the preset sampling frequency is within a preset error range, gradually increasing the preset sampling frequency until the measured frequency error value between the actual sampling frequency and the preset sampling frequency exceeds the preset error range, and taking the current preset sampling frequency as the maximum sampling frequency of the sensor; acquiring a to-be-identified limb signal of a first user, which is acquired by the sensor in the mobile terminal at the maximum sampling frequency;
the input unit is used for inputting the limb signal to be recognized into the trained recognition model;
and the sending unit is used for acquiring the recognition model, recognizing the limb signal to be recognized according to the trained network parameters, outputting a recognition result and sending the recognition result to the mobile terminal.
7. The apparatus of claim 6, wherein the building unit comprises:
the acquiring subunit is used for acquiring a plurality of first limb signals of healthy human bodies and a plurality of second limb signals of patients with limb tremor symptoms;
the manufacturing subunit is used for respectively manufacturing samples of the plurality of first limb signals and the plurality of second limb signals according to a preset format to obtain a training set comprising a plurality of training samples;
a construction subunit configured to construct the recognition model;
the training subunit is used for inputting the training set into the convolutional neural network of the recognition model to obtain the forward output of the convolutional neural network;
and the updating subunit is used for updating the weight and the bias of the convolutional neural network by using a back propagation neural network algorithm to obtain the trained recognition model and storing the network parameters of the trained recognition model.
8. A non-volatile storage medium for a computer, the storage medium comprising a stored program, wherein the program when executed controls an apparatus in which the storage medium is located to perform the steps of any one of claims 1 to 5 when the program instructions in the server are loaded and executed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811422696.1A CN109276255B (en) | 2018-11-27 | 2018-11-27 | Method and device for detecting tremor of limbs |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811422696.1A CN109276255B (en) | 2018-11-27 | 2018-11-27 | Method and device for detecting tremor of limbs |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109276255A CN109276255A (en) | 2019-01-29 |
| CN109276255B true CN109276255B (en) | 2023-02-28 |
Family
ID=65172826
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811422696.1A Active CN109276255B (en) | 2018-11-27 | 2018-11-27 | Method and device for detecting tremor of limbs |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109276255B (en) |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109949932A (en) * | 2019-03-28 | 2019-06-28 | 南开大学 | A kind of judgment method of the essential tremor extent based on machine learning |
| CN110037683B (en) * | 2019-04-01 | 2020-12-29 | 上海数创医疗科技有限公司 | Improved convolutional neural network for recognizing heart rhythm type and training method thereof |
| CN110037684A (en) * | 2019-04-01 | 2019-07-23 | 上海数创医疗科技有限公司 | Device based on the identification rhythm of the heart type for improving convolutional neural networks |
| CN110522456A (en) * | 2019-09-26 | 2019-12-03 | 安徽中医药大学 | A kind of WD based on deep learning trembles conditions of patients self-evaluating system |
| CN110522455A (en) * | 2019-09-26 | 2019-12-03 | 安徽中医药大学 | A kind of WD based on deep learning trembles grade appraisal procedure |
| CN111292844B (en) * | 2020-01-21 | 2023-03-21 | 桂林医学院附属医院 | Parkinson disease condition monitoring system |
| CN112075940A (en) * | 2020-09-21 | 2020-12-15 | 哈尔滨工业大学 | Tremor detection system based on bidirectional long-time and short-time memory neural network |
| CN113282203A (en) * | 2021-04-30 | 2021-08-20 | 深圳市联谛信息无障碍有限责任公司 | Interface switching method and device for user with tremor of limbs and electronic equipment |
| CN113229798B (en) * | 2021-05-18 | 2023-08-22 | 平安科技（深圳）有限公司 | Model migration training method, device, computer equipment and readable storage medium |
| CN113240018B (en) * | 2021-05-19 | 2023-02-03 | 哈尔滨医科大学 | Hand-drawn graph classification method and system based on error back propagation algorithm |
| CN113855570B (en) * | 2021-09-30 | 2023-07-14 | 平安科技（深圳）有限公司 | Parkinson's disease medicine taking reminding method, system, electronic equipment and storage medium |
| CN116757534B (en) * | 2023-06-15 | 2024-03-15 | 中国标准化研究院 | Intelligent refrigerator reliability analysis method based on neural training network |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017167518A1 (en) * | 2016-03-31 | 2017-10-05 | Fotonation Limited | Biometric recognition system |
| WO2017192028A1 (en) * | 2016-05-03 | 2017-11-09 | Universiti Sains Malaysia | Tremor motion quantifying method and system |
| CN107958257A (en) * | 2017-10-11 | 2018-04-24 | 华南理工大学 | A kind of Chinese traditional medicinal materials recognition method based on deep neural network |
| CN108197648A (en) * | 2017-12-28 | 2018-06-22 | 华中科技大学 | A kind of Fault Diagnosis Method of Hydro-generating Unit and system based on LSTM deep learning models |
| CN108875608A (en) * | 2018-06-05 | 2018-11-23 | 合肥湛达智能科技有限公司 | A kind of automobile traffic signal recognition method based on deep learning |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2273484B1 (en) * | 2004-04-30 | 2019-08-28 | Hillcrest Laboratories, Inc. | Methods and devices for removing unintentional movement in free space pointing devices |
| US10656695B2 (en) * | 2015-03-25 | 2020-05-19 | Koninklijke Philips N.V. | Health wearable that automatically changes sensor reading timings |
-
2018
- 2018-11-27 CN CN201811422696.1A patent/CN109276255B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017167518A1 (en) * | 2016-03-31 | 2017-10-05 | Fotonation Limited | Biometric recognition system |
| WO2017192028A1 (en) * | 2016-05-03 | 2017-11-09 | Universiti Sains Malaysia | Tremor motion quantifying method and system |
| CN107958257A (en) * | 2017-10-11 | 2018-04-24 | 华南理工大学 | A kind of Chinese traditional medicinal materials recognition method based on deep neural network |
| CN108197648A (en) * | 2017-12-28 | 2018-06-22 | 华中科技大学 | A kind of Fault Diagnosis Method of Hydro-generating Unit and system based on LSTM deep learning models |
| CN108875608A (en) * | 2018-06-05 | 2018-11-23 | 合肥湛达智能科技有限公司 | A kind of automobile traffic signal recognition method based on deep learning |
Non-Patent Citations (1)
| Title |
|---|
| 基于大规模医疗数据的病症识别与预警;朱凤超;《中国优秀博硕士学位论文全文数据库（硕士） 医药卫生科技辑》;20170215(第2期);摘要、正文第5-66页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109276255A (en) | 2019-01-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109276255B (en) | Method and device for detecting tremor of limbs | |
| US20190365332A1 (en) | Determining wellness using activity data | |
| KR101690649B1 (en) | Activity classification in a multi-axis activity monitor device | |
| Lester et al. | A practical approach to recognizing physical activities | |
| US10636524B2 (en) | Method and system for optimized wake-up strategy via sleeping stage prediction with recurrent neural networks | |
| CN109363659B (en) | Heart rate monitoring method and device based on deep learning and storage medium | |
| CN109684118A (en) | Detection method, device, equipment and the computer readable storage medium of abnormal data | |
| He et al. | A low power fall sensing technology based on FD-CNN | |
| WO2017050140A1 (en) | Method for recognizing a human motion, method for recognizing a user action and smart terminal | |
| WO2010083562A1 (en) | Activity detection | |
| CN106210269A (en) | A kind of human action identification system and method based on smart mobile phone | |
| KR101967342B1 (en) | An exercise guide system by using wearable device | |
| CN112057066A (en) | Heart rate detection method, wearable device and computer storage medium | |
| CN116821809B (en) | Vital sign data acquisition system based on artificial intelligence | |
| CN112690761B (en) | Sleep state detection method, device, equipment and computer readable medium | |
| Khokhlov et al. | Design of activity recognition systems with wearable sensors | |
| Aiolli et al. | ClimbTheWorld: Real-time stairstep counting to increase physical activity | |
| Navarathna et al. | Wearable device based activity recognition and prediction for improved feedforward control | |
| CN111652138A (en) | Face recognition method, device and equipment for wearing mask and storage medium | |
| CN110598599A (en) | Method and device for detecting abnormal gait of human body based on Gabor atomic decomposition | |
| CN113689660B (en) | Safety early warning method of wearable device and wearable device | |
| CN110866672A (en) | Data processing method, device, terminal and medium | |
| US20220151511A1 (en) | System, apparatus and method for activity classification for a watch sensor | |
| KR101483218B1 (en) | Activity diagnosis apparatus | |
| Varma et al. | Health assessment for office workers by tracking the wrist motion centric activity |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |